1.. _pagemap:
2
3=============================
4Examining Process Page Tables
5=============================
6
7pagemap is a new (as of 2.6.25) set of interfaces in the kernel that allow
8userspace programs to examine the page tables and related information by
9reading files in ``/proc``.
10
11There are four components to pagemap:
12
13 * ``/proc/pid/pagemap``.  This file lets a userspace process find out which
14   physical frame each virtual page is mapped to.  It contains one 64-bit
15   value for each virtual page, containing the following data (from
16   ``fs/proc/task_mmu.c``, above pagemap_read):
17
18    * Bits 0-54  page frame number (PFN) if present
19    * Bits 0-4   swap type if swapped
20    * Bits 5-54  swap offset if swapped
21    * Bit  55    pte is soft-dirty (see
22      :ref:`Documentation/admin-guide/mm/soft-dirty.rst <soft_dirty>`)
23    * Bit  56    page exclusively mapped (since 4.2)
24    * Bit  57    pte is uffd-wp write-protected (since 5.13) (see
25      :ref:`Documentation/admin-guide/mm/userfaultfd.rst <userfaultfd>`)
26    * Bits 58-60 zero
27    * Bit  61    page is file-page or shared-anon (since 3.5)
28    * Bit  62    page swapped
29    * Bit  63    page present
30
31   Since Linux 4.0 only users with the CAP_SYS_ADMIN capability can get PFNs.
32   In 4.0 and 4.1 opens by unprivileged fail with -EPERM.  Starting from
33   4.2 the PFN field is zeroed if the user does not have CAP_SYS_ADMIN.
34   Reason: information about PFNs helps in exploiting Rowhammer vulnerability.
35
36   If the page is not present but in swap, then the PFN contains an
37   encoding of the swap file number and the page's offset into the
38   swap. Unmapped pages return a null PFN. This allows determining
39   precisely which pages are mapped (or in swap) and comparing mapped
40   pages between processes.
41
42   Efficient users of this interface will use ``/proc/pid/maps`` to
43   determine which areas of memory are actually mapped and llseek to
44   skip over unmapped regions.
45
46 * ``/proc/kpagecount``.  This file contains a 64-bit count of the number of
47   times each page is mapped, indexed by PFN.
48
49The page-types tool in the tools/vm directory can be used to query the
50number of times a page is mapped.
51
52 * ``/proc/kpageflags``.  This file contains a 64-bit set of flags for each
53   page, indexed by PFN.
54
55   The flags are (from ``fs/proc/page.c``, above kpageflags_read):
56
57    0. LOCKED
58    1. ERROR
59    2. REFERENCED
60    3. UPTODATE
61    4. DIRTY
62    5. LRU
63    6. ACTIVE
64    7. SLAB
65    8. WRITEBACK
66    9. RECLAIM
67    10. BUDDY
68    11. MMAP
69    12. ANON
70    13. SWAPCACHE
71    14. SWAPBACKED
72    15. COMPOUND_HEAD
73    16. COMPOUND_TAIL
74    17. HUGE
75    18. UNEVICTABLE
76    19. HWPOISON
77    20. NOPAGE
78    21. KSM
79    22. THP
80    23. OFFLINE
81    24. ZERO_PAGE
82    25. IDLE
83    26. PGTABLE
84
85 * ``/proc/kpagecgroup``.  This file contains a 64-bit inode number of the
86   memory cgroup each page is charged to, indexed by PFN. Only available when
87   CONFIG_MEMCG is set.
88
89Short descriptions to the page flags
90====================================
91
920 - LOCKED
93   The page is being locked for exclusive access, e.g. by undergoing read/write
94   IO.
957 - SLAB
96   The page is managed by the SLAB/SLOB/SLUB/SLQB kernel memory allocator.
97   When compound page is used, SLUB/SLQB will only set this flag on the head
98   page; SLOB will not flag it at all.
9910 - BUDDY
100    A free memory block managed by the buddy system allocator.
101    The buddy system organizes free memory in blocks of various orders.
102    An order N block has 2^N physically contiguous pages, with the BUDDY flag
103    set for and _only_ for the first page.
10415 - COMPOUND_HEAD
105    A compound page with order N consists of 2^N physically contiguous pages.
106    A compound page with order 2 takes the form of "HTTT", where H donates its
107    head page and T donates its tail page(s).  The major consumers of compound
108    pages are hugeTLB pages
109    (:ref:`Documentation/admin-guide/mm/hugetlbpage.rst <hugetlbpage>`),
110    the SLUB etc.  memory allocators and various device drivers.
111    However in this interface, only huge/giga pages are made visible
112    to end users.
11316 - COMPOUND_TAIL
114    A compound page tail (see description above).
11517 - HUGE
116    This is an integral part of a HugeTLB page.
11719 - HWPOISON
118    Hardware detected memory corruption on this page: don't touch the data!
11920 - NOPAGE
120    No page frame exists at the requested address.
12121 - KSM
122    Identical memory pages dynamically shared between one or more processes.
12322 - THP
124    Contiguous pages which construct transparent hugepages.
12523 - OFFLINE
126    The page is logically offline.
12724 - ZERO_PAGE
128    Zero page for pfn_zero or huge_zero page.
12925 - IDLE
130    The page has not been accessed since it was marked idle (see
131    :ref:`Documentation/admin-guide/mm/idle_page_tracking.rst <idle_page_tracking>`).
132    Note that this flag may be stale in case the page was accessed via
133    a PTE. To make sure the flag is up-to-date one has to read
134    ``/sys/kernel/mm/page_idle/bitmap`` first.
13526 - PGTABLE
136    The page is in use as a page table.
137
138IO related page flags
139---------------------
140
1411 - ERROR
142   IO error occurred.
1433 - UPTODATE
144   The page has up-to-date data.
145   ie. for file backed page: (in-memory data revision >= on-disk one)
1464 - DIRTY
147   The page has been written to, hence contains new data.
148   i.e. for file backed page: (in-memory data revision >  on-disk one)
1498 - WRITEBACK
150   The page is being synced to disk.
151
152LRU related page flags
153----------------------
154
1555 - LRU
156   The page is in one of the LRU lists.
1576 - ACTIVE
158   The page is in the active LRU list.
15918 - UNEVICTABLE
160   The page is in the unevictable (non-)LRU list It is somehow pinned and
161   not a candidate for LRU page reclaims, e.g. ramfs pages,
162   shmctl(SHM_LOCK) and mlock() memory segments.
1632 - REFERENCED
164   The page has been referenced since last LRU list enqueue/requeue.
1659 - RECLAIM
166   The page will be reclaimed soon after its pageout IO completed.
16711 - MMAP
168   A memory mapped page.
16912 - ANON
170   A memory mapped page that is not part of a file.
17113 - SWAPCACHE
172   The page is mapped to swap space, i.e. has an associated swap entry.
17314 - SWAPBACKED
174   The page is backed by swap/RAM.
175
176The page-types tool in the tools/vm directory can be used to query the
177above flags.
178
179Using pagemap to do something useful
180====================================
181
182The general procedure for using pagemap to find out about a process' memory
183usage goes like this:
184
185 1. Read ``/proc/pid/maps`` to determine which parts of the memory space are
186    mapped to what.
187 2. Select the maps you are interested in -- all of them, or a particular
188    library, or the stack or the heap, etc.
189 3. Open ``/proc/pid/pagemap`` and seek to the pages you would like to examine.
190 4. Read a u64 for each page from pagemap.
191 5. Open ``/proc/kpagecount`` and/or ``/proc/kpageflags``.  For each PFN you
192    just read, seek to that entry in the file, and read the data you want.
193
194For example, to find the "unique set size" (USS), which is the amount of
195memory that a process is using that is not shared with any other process,
196you can go through every map in the process, find the PFNs, look those up
197in kpagecount, and tally up the number of pages that are only referenced
198once.
199
200Exceptions for Shared Memory
201============================
202
203Page table entries for shared pages are cleared when the pages are zapped or
204swapped out. This makes swapped out pages indistinguishable from never-allocated
205ones.
206
207In kernel space, the swap location can still be retrieved from the page cache.
208However, values stored only on the normal PTE get lost irretrievably when the
209page is swapped out (i.e. SOFT_DIRTY).
210
211In user space, whether the page is present, swapped or none can be deduced with
212the help of lseek and/or mincore system calls.
213
214lseek() can differentiate between accessed pages (present or swapped out) and
215holes (none/non-allocated) by specifying the SEEK_DATA flag on the file where
216the pages are backed. For anonymous shared pages, the file can be found in
217``/proc/pid/map_files/``.
218
219mincore() can differentiate between pages in memory (present, including swap
220cache) and out of memory (swapped out or none/non-allocated).
221
222Other notes
223===========
224
225Reading from any of the files will return -EINVAL if you are not starting
226the read on an 8-byte boundary (e.g., if you sought an odd number of bytes
227into the file), or if the size of the read is not a multiple of 8 bytes.
228
229Before Linux 3.11 pagemap bits 55-60 were used for "page-shift" (which is
230always 12 at most architectures). Since Linux 3.11 their meaning changes
231after first clear of soft-dirty bits. Since Linux 4.2 they are used for
232flags unconditionally.
233